Highly efficient prismatic perovskite solar cells

Jiang Huang a, Siheng Xiang a, Junsheng Yu *a and Chang-Zhi Li *b
aState Key Laboratory of Electronic Thin Films and Integrated Devices, School of Optoelectronic Science and Engineering, University of Electronic Science and Technology of China (UESTC), Chengdu 610054, P. R. China. E-mail: jsyu@uestc.edu.cn
bMOE Key Laboratory of Macromolecular Synthesis and Functionalization, State Key Laboratory of Silicon Materials, Department of Polymer Science and Engineering, Zhejiang University, Hangzhou 310027, P. R. China. E-mail: czli@zju.edu.cn

Received 4th September 2018 , Accepted 22nd November 2018

First published on 22nd November 2018

Lead halide perovskites, such as methylammonium lead iodide (MAPbI3) can reach near 100% internal quantum efficiency in single solar cells, but they still encounter significant thermodynamic losses in photon energy to offset device photovoltage and performance. Herein, a novel prismatic perovskite solar cell with light trapping configuration, namely, Prim PVSC, is designed to mitigate such losses in devices, through modulating the pathway of light inside series cells, wherein incident high-to-low energy photons are separately captured by four horizontally aligned MAPbIxBr3−x subcells with varied bandgaps. This newly designed PVSC has remarkably led to a record open circuit voltage of 5.3 V in four series devices and power conversion efficiency of 21.3%, which could provide a new way to break the performance bottleneck of perovskites. Practically, this type of device architecture could also be applied in flexible modules for large-area application.

Broader context

Solar photon-to-electron conversion with lead halide perovskites (PVKs) is an emerging and promising clean energy solution that has attracted significant interest from the research community. In recent years, photovoltaics consisting of PVK absorbers have rapidly improved their power conversion efficiencies (PCEs), with the performance of the best lab cell surpassing that of polycrystalline silicon. A significant portion of energy loss routinely occurs in the single junction PVK cells, where photon energies above the optical bandgap of absorbers (EqVoc) are lost, known as carrier thermalization, resulting in the thermodynamic losses in the Shockley–Queisser limit. In this work, prismatic perovskite solar cells (Prim PVSC) are developed to mitigate the thermodynamic losses of traditional single junction PVK cells. By guiding the flow of light inside the Prim PVSC, the solar photons with high-to-low energy are separately captured by the four subcells with varied, yet matched, bandgaps of MAPbIxBr3−x films. This is the first Prim PVSC with four series subcells to generate a record Voc of 5.3 V, leading to a high power conversion efficiency of 21.3%, providing a new method for potentially breaking the performance bottleneck of PVSCs.


Lead halide perovskites (PVKs) have attracted significant interest from the research community as photon absorbers for photovoltaic applications in recent years, due to their distinct advantages of small exciton binding energies, high absorption coefficient, and long carrier diffusion lengths.1,2 Since the first perovskite solar cells (PVSC) were reported by Kojima et al.,3 both the solution process4–6 and the thermal evaporation method7–9 have been proposed to further improve the crystallinity and device performance. To date, the power conversion efficiencies (PCE) have been rapidly expanded to surpass 23%4 with the advancements in materials, processing and device architecture.5,6,10 The high efficiency, together with the prospect of cheap precursors, makes the PVSC one of the most promising candidates for a new form of clean energy.11,12

So far, the most investigated perovskite, CH3NH3PbI3 (MAPbI3), can already reach near 100% internal quantum efficiency in single junction PVSCs,6,13,14 which means that to further improve the efficiency of such cells would be particularly difficult through the optimization of single absorber layers. However, the single junction PVSC still encounters certain energy losses due to the relatively low open circuit voltage (Voc) versus energy bandgap. To further improve PVSC efficiency, new PVK materials have been designed; for instance, the MA ions could be replaced by other cations, i.e. amidinium CH(NH2)2+(FA)15 or metal ions of Cs16 or Rb.17–19 Also, bandgap tuning of MAPbI3 can be achieved via substituting halide anions with Br ions, yielding MAPbI3−xBrx.20 Notably, by tuning the ratio of I and Br ions, the band gap of MAPbIxBr3−x can continuously be adjusted from 1.5 eV to 2.3 eV.21 Furthermore, the MA/FA and Br/I-ratios can be simultaneously changed to form mixed perovskites, wherein the state-of-the-art PVKs have some of the iodide replaced by bromide and the majority of MA replaced by FA.15,22 Interfacial materials such as fullerene derivatives,23–25 non-fullerenes26,27 and metal oxides28 as electron transporting layers, as well as nickel oxide (NiO),29 aryl amine molecules30–33 and polymers34,35 as hole transporting layers have been used in primary charge extraction, with the additional function of defect passivation for perovskites. Regarding device structure, tandem devices that combine cost-effective PVSCs with low-bandgap silicon36 or CIGS37 solar cells are promising for deploying PVSCs into communities to work complementarily with the established solar cell technologies.38,39 Moreover, the PVSC/organic bulk heterojunction hybrid devices have been proposed for further improving the current efficiency ceilings.40,41

Taking the optical bandgaps of active layers into consideration, the MAPbI3 active layer with 1.5 eV optical bandgap can harvest most of the solar photons from 350 nm to 800 nm (photon energy E = ħω, from 3.5 eV to 1.5 eV). However, the MAPbI3 device only exhibits a photovoltage Voc about 1 V. Those photon energies above the optical bandgap of absorbers (EqVoc) are lost, known as carrier thermalization, resulting in thermodynamic losses.15 According to the Shockley–Queisser model,42 thermodynamic loss could be mitigated by reducing the deficit between the bandgap energy and the electron–hole quasi-Fermi-level splitting, and also decreasing carrier thermalization losses to best develop the efficiency of solar cells.43 The initial consideration for tandem structures is to separately convert high energy photons in the front cells with large band gaps to generate high Voc, and low-energy photons harvested by rear cells with narrow band gap materials through a stacking tandem device.44 Solar cells with spectrum-splitting photonic configurations45,46 or folded reflective tandem structures have also been explored.47 However, the actual application of large band gap PVKs (i.e., MAPbBr3 or MAPbCl3) as front cells in the tandem device architecture is rarely reported.44,48 Alternatively, several works use MAPbI3 PVSC as an add-on to silicon or other commercial solar cells where the PVSC is the front cell using both 2 and 4 terminal tandem architectures to realize high Voc and efficiency.49–51 To date, effective pathways to efficiently mitigate high thermodynamic losses of PVSCs, as well as make full use of large bandgap PVK materials, are still lacking.39

Herein, a new form of solar cell architecture with light trapping configuration, namely Prim PVSC, is realized by the series connection of subcells with different bandgaps in horizontal alignment, wherein incident photons with different energies are subsequently harvested by four different MAPbIxBr3−x subcells to minimize thermodynamic losses and improve the photovoltages of PVSCs. By modulating the pathway of incident light based on a light trapping architecture, the specific high-to-low energy photons were separately captured by the individual subcells consisting of absorbers with varied bandgaps, i.e. MAPbBrxI3−x (x = 3, 2, 1 and 0). Four subcells generated Voc from 1.54 V to 1.15 V, respectively, thus, a record Voc of 5.3 V in the four series subcells was obtained. This newly designed Prim PVSC remarkably leads to PCE enhancement from 18% of the neat MAPbI3 device to 21.3% (Prim PVSC), accounting for 18% thermodynamic loss mitigation of the neat MAPbI3 device. Overall, through loss mitigation and light trapping of the new device architecture, Prim PVSC allows the simultaneous achievement of high Voc and PCE, which would provide a new way to break the performance bottleneck of perovskites.

Device configuration

Fig. 1a and b show the energy level diagrams and extinction coefficients of MAPbI3−xBrx PVK films. The optical energy bandgaps of MAPbBr3, MAPbIBr2, MAPbI2Br, and MAPbI3 films were determined from extinction coefficients to be 2.3 eV, 2.05 eV, 1.8 eV and 1.5 eV, respectively, which agree well with a previous report.21Fig. 1b also shows that the large bandgap of MAPbBr3 absorbs solar light up to 550 nm; MAPbIBr2 and MAPbI2Br have absorption ranges up to 640 nm and 720 nm, respectively; the MAPbI3 film can absorb up to 800 nm, which can match with the different photon energies. Fig. 1c and d show the device structures of the single and modular Prim PVSC in this work. The device configuration is indium tin oxide (ITO)/NiO/MAPbI3−xBrx (x = 0, 1, 2 or 3)/electron transporting layer (ETL) (ICBA, PCBM or ICBA:PCBM)/Al, including four subcells placed in the horizontal plane. Ag film (100 nm) is thermal evaporated on the glass substrate as the opposite reflector. The oblique sunlight firstly locates at Cell 1, where the high-energy photons are absorbed by the MAPbBr3 layer. Then, the sunlight is reflected by the opposite reflector to Cell 2 and absorbed by the MAPbIBr2 active layer. Similarly, the incident solar light was successively absorbed by the other two subcells. Four subcells sequentially absorb incident photons ranging from high to low energy, which could generate higher Voc in subcells 1–3 as compared to MAPbI3 PVSC. Also, it shows that the incident light would be trapped in the Prim cells, which is beneficial for the best photon-harvesting. It is worth noting that this device is feasibly constructed by using the thermal evaporation method, which allows the precise control of the position and alignment of active layers toward the top electrodes.
image file: c8ee02575d-f1.tif
Fig. 1 Energy diagrams, light absorption spectra of active materials and device structures in this work. (a) Energy level diagrams of PVK solar cells; (b) extinction coefficients of PVK films including MAPbBr3, MAPbIBr2, MAPbI2Br, and MAPbI3. (c) Device configuration of Prim PVSC including 4 subcells. (d) Prim PVSC module structure with the path of solar light.

To construct Prim PVSCs, we employed the four-source co-evaporation method to fabricate MAPbI3−xBrx active films, as shown in Fig. 2a. In this work, MAPbBr3, MAPbIBr2, MAPbI2Br and MAPbI3 active films were evaporated through a shadow mask to form four horizontal layers from Cell 1 to Cell 4, respectively. After that, the substrates were subjected to DMF[thin space (1/6-em)]:[thin space (1/6-em)]CB (1[thin space (1/6-em)]:[thin space (1/6-em)]20 v/v) co-solvent vapor treatment for 10 min52 to improve the crystallinity of the PVK films. As shown in Fig. S1, ESI, the solvent vapor treatment could promote the uniform perovskite films to form larger crystalline grains and rougher surfaces, which benefit charge transportation and light scattering. The PVK films were then thermally annealed on the hot plate to remove solvent residues, as shown in Fig. 2b. Subsequently, drops of ETLs of ICBA and PCBM solutions were dropped onto the film surface of MAPbBr3 and MAPbI3, respectively; the nearby drops gradually came together during the evaporation of the solvent on the hotplate. The ETL on the MAPbBr3 and MAPbI3 active films were pure ICBA and PCBM layers, respectively, while the ETL on the MAPbIBr2 and MAPbI2Br films were mixtures of ICBA and PCBM, as shown in Fig. 2c. After dip-coating the ETLs, the patterned Al electrodes were evaporated through a mask onto the ETLs to form the connected Prim device, whose top-view structure is shown in Fig. 2d. The four subcells were connected in series by the Al electrode and the patterned ITO anode. The distance between the subcells is 50 μm, defined by the separated Al electrodes; a wider gap could lead to light leakage, which would reduce light absorption and the overall device efficiency.

image file: c8ee02575d-f2.tif
Fig. 2 The process of device fabrication of serial Prim PVK solar cells. (a) Four-source co-evaporation of MAPbIxBr3−x films. (b) Solvent vapor treatment of PVK films. (c) Dip coating of the electron transporting layer of ICBA and PCBM on PVK films. (d) Top view of the fabricated Prim solar cell. The four devices are connected in series using the Al electrode.

Results and discussion

To ensure the quality of PVK films using the four-source thermal evaporation method, the crystallinity of four active films were investigated. Fig. S2 (ESI) exhibits the X-ray diffraction (XRD) patterns monitored in the 2θ range of 27.5–30.5° for MAPbI3−xBrx films. It can be seen that two peaks of MAPbI3 film are located at 28.1° and 28.3° which were indexed to the (004) and (220) planes for the tetragonal I4/mcm phase.3,21 When x ≥1 for MAPbI3−xBrx, the diffracted peak of the (004) plane disappears into a single peak corresponding to (200), due to the increased symmetry. This means that the tetragonal I4/mcm phase of MAPbI3 turns to the cubic Pm[3 with combining macron]m phase of MAPbI2Br, MAPbIBr2, and MAPbBr3. The systematic shift of the (200)c peak toward higher 2θ degrees with further introduction of Br ions into MAPbI3−xBrx is because the gradual substitution of the larger I atoms with the smaller Br atoms decreases the lattice spacing.21

In order to investigate the effects of the ETLs, PL spectra of MAPbI3−xBrx and MAPbI3−xBrx/ETL (ICBA, PCBM or ICBA: PCBM) were recorded and are presented in Fig. 3, which show that the energy of PL emission depends on the halide components of PVK films, ranging across the visible spectrum. The PL peak positions of MAPbBr3, MAPbIBr2, MAPbI2Br and MAPbI3 films show that the energies of the PL peaks are always red-shifted with respect to the onset of the absorption edge. The PL intensities of NiO/MAPbI3−xBrx films were quenched by 70% and 83% on using ICBA or PCBM ETLs. It can be seen that the strong PL quenching originated from the enhanced exciton dissociation at the interface between the PVK film and ETL.20 Moreover, additional long-term solvent annealing was adopted to post-treat NiO/PVK/ETL films at low temperature, which has proved that ETL molecules can gradually penetrate the PVK film to mend both surface and bulk defects for improving the Jsc and FF of PVSCs.20

image file: c8ee02575d-f3.tif
Fig. 3 Normalized PL spectra of PVK films with and without ETL.

To investigate the halide segregation in mixed halide perovskite, we measured PL spectra as a function of time. As shown in Fig. S3 (ESI), the PL peak of the MAPbI2Br film has a slight red-shift without the generation of obvious subpeaks after illumination for 1 and 2 minutes under 457 nm light, which is similar to previously reported work.53 The PL features of the MAPbIBr2 film also showed a slight red-shift after illumination, and a new low-energy PL emission peak appeared at 1.75 eV, ascribed to the photo-induced halide segregation observed by Hoke et al.53 However, the intensity of this low-energy peak is much weaker than that of the main peak of 2.0 eV, indicating a rather small portion. The photo-induced halide segregation could result in a reduction in the electronic bandgap and quasi-Fermi level splitting.53

Optical modeling was conducted by using the transfer matrix method (TMM) to investigate the optical distribution and absorption of incident light in the studied devices. All the complex refraction indexes (n + ik) of studied materials for simulation were measured by spectroscopic ellipsometry and are presented in Fig. S4 and S5 (ESI). To match the current density of the subcells in series is a critical challenge due to the overall photocurrent of series cells being determined by the minimum current value of each subcell. To achieve this, we proposed a simple method of matching the light absorption and photocurrent of four subcells. Firstly, the average short-circuit current density Jsc of four subcells was estimated to be around 5–7 mA cm−2. Secondly, the thickness dependent Jsc of the top two cells were calculated (Fig. 4a), and the min. The Jsc1&2 curve was extracted in Fig. 4b, which shows that the current increased monotonously with the thicker PVK films of the top two cells along with the dark line. Thirdly, the thickness-dependent Jsc and min. Jsc1&2&3 curves of Cells 1 to 3 are presented in Fig. 4c and d. It can be seen that the optional thicknesses of Cells 1 to 3 are also limited at the dark line in Fig. 4d. Finally, thickness dependent Jsc and the min. The Jsc1&2&3&4 curves of Cells 1 to 4 were obtained in Fig. 4e and f. Notably, the matching current of min. Jsc1&2&3&4 did not increase monotonously and a local maximum value of about 5.7 mA cm−2 was obtained. According to this strategy, the optimal thicknesses of PVK films in four subcells were chosen as 93 nm of MAPbBr3, 122 nm of MAPbIBr2, 160 nm of MAPbI2Br, and 340 nm of MAPbI3.

image file: c8ee02575d-f4.tif
Fig. 4 Thickness dependence of the overall and separate photocurrents (Jsc) on the active perovskite layers of four subcells. (a) and (b) are Jsc optimizations for Cell 1 and Cell 2. (c) and (d) are Jsc optimizations for Cells 1–2 and Cell 3. (e) and (f) are also Jsc optimizations for Cells 1–2–3 and Cell 4.

We further employed panchromatic optical field simulation to investigate the light absorption, incident and reflected light spectra in the device configuration of ITO/NiO/PVK (93 nm of MAPbBr3, 122 nm of MAPbIBr2, 160 nm of MAPbI2Br, and 340 nm of MAPbI3)/ETL/Al. Fig. S6 (ESI) shows that Cell 1 only absorbed ∼70% of incident light I1 in the range of 400 nm to 550 nm, and reflected light R1 could participate in the incident light I2 to generate more photocurrent for Cell 2. The reflection rate R2 contained 40% of solar light for Cell 3 in the range of 500 nm to 650 nm. The reflection rate R3 contained 20% of solar light for Cell 4 in the range of 600 nm to 750 nm. After Cell 4, with 340 nm MAPbI3 film, absorbed the remaining solar light, the reflected light R4 was equal to zero in the range of 350 nm to 750 nm, and the remaining near-infrared part could be used as the incident light for other solar cells, i.e., Si or Ge cells.49,54Fig. 5 clearly shows the charge generation in the four subcells. The charge generation rates in Cells 1 and 2 have some overlap in the range of 400 nm to 550 nm. Likewise, there was some kind of overlap between Cells 2 and 3, and Cells 3 and 4. From the energy conversion point of view, this kind of light absorption overlap means that there are some thermodynamic losses for the overall device, which could be avoided by using higher bandgap subcells.

image file: c8ee02575d-f5.tif
Fig. 5 Charge generation rate in the active layers of serial Prim PVK cells. (a) to (d) refer to MAPbBr3, MAPbIBr2, MAPbI2Br, and MAPbI3 layers in Cells 1 to 4, respectively.

Prior to the construction of Prim PVSCs, the photovoltaic performances of the individual single-junction PVSCs were tested. Fig. S7a (ESI) shows the current density versus voltage (JV) curves of individual single-junction PVSCs and the relevant photovoltaic parameters are summarized in Table S1 (ESI). Cell 1 with the 93 nm MAPbBr3 layer has a high Voc of 1.55 V, Jsc of 5.5 mA cm−2, and FF of 0.77, leading to PCE of 6.44%, where the Voc is slightly lower than the reported 1.6 V.20 The high Voc originates from the ICBA as the ETL, which has a higher lying lowest unoccupied molecular orbital (LUMO) level of −3.7 eV as compared to PCBM (−3.9 eV). The LUMO level of ICBA matches well with the conduction band of the MAPbBr3 layer, which could increase the built-in voltage.20 Cell 2 with the single 122 nm MAPbIBr2 active layer has Voc of 1.45 V, Jsc of 10.14 mA cm−2, FF of 0.76, and PCE of 11.2%. Cell 3 with the single 160 nm MAPbI2Br active layer has Voc of 1.26 V, Jsc of 15.8 mA cm−2, FF of 0.77, and PCE of 15.33%. The Voc of Cell 2 and Cell 3 are also enhanced by using the ICBA:PCBM mixed ETL as compared with the reported mixed halide perovskite.55,56 The maximum power point tracking (MPPT) profiles of MAPbI2Br and MAPbIBr2 PVSCs are presented in Fig. S8 (ESI). It can be seen that the device PCEs slightly decreased in the first 4 minutes, and then became steady for both MAPbI2Br and MAPbIBr2 devices; the photocurrents and fill factors of both devices were relatively stable. The efficiency roll-off under continuous light illumination could be attributed to the halide segregation of the mixed bromide iodide perovskite,53 which is also evidenced from the red-shift and low-energy sub-peak of the PL spectrum for the mixed bromide iodide perovskite after light illumination in Fig. S3 (ESI). Cell 4, with a single 340 nm MAPbI3 active layer has Voc of 1.15 V, Jsc of 20.75 mA cm−2, FF of 0.78, and PCE of 18.6%. The Voc of the MAPbI3-based device is relatively higher than the solution processed PVSCs,57 and the PCE is also among the state-of-the-art performance of the thermally evaporated PVSCs.58

This method differs from the conventional thermal evaporation process in that we treated the PVK film with the solvent vapor and added solution processed ETL to the PVK film. The crystallinity of the PVK film was enhanced in the first case, which was beneficial for charge transport and good film morphology.52 In the second case, it was proved that the solution processed ETL could infiltrate the grain boundaries of the PVK film and promote electron extraction and collection, which could ensure the overall improvement in Voc, Jsc and FF performance.20,59 Fig. S7b (ESI) presents the external quantum efficiency (EQE) of individual PVSCs. It shows that the EQE value of Cells 1 and 2 are relatively low due to the inadequate absorption of solar light at the relatively thin active layer of about 100 nm, while the EQE values of Cell 3 and Cell 4 are very high. In particular, the maximum EQE value of Cell 4 reached about 90%, due to the adequate absorption of solar light and also near 100% internal quantum efficiency.

Fig. 6a shows the JV curves of Prim PVSCs, and their photovoltaic performances are summarized in Table 1. Fittings of JV curves for extracting the equivalent circuit parameters of PVSCs are shown in Fig. S9 (ESI). Interestingly, for this series of Prim solar cells, the performance of each subcell could be simply tested by the direct contact of the anode and electrode. The four subcells in the Prim connected structure have similar Jsc of about 5.38–5.53 mA cm−2, almost equal to the calculated Jsc of about 5.7 mA cm−2 based on optical simulation and the integrated current density of JEQE from the EQE curves. The almost equal current is important in the series connection designation. Also, the Voc and FF parameters do not change much for four subcells. However, Cell 2, Cell 3 and Cell 4 have reduced PCEs of 6.07%, 5.43% and 4.81% due to reduced light absorption and Jsc. The photovoltaic performance of the series connected Prim cell was tested via the outer anode and electrode. The overall Jsc and FF of the Prim cell were slightly reduced to 5.35 mA cm−2 and 0.75 due to some charge losses in the Prim cell and cumulative series resistance of the four subcells. Encouragingly, the Voc of the Prim cell was as high as 5.29 V, which is the highest value of four series devices in literature, leading to the PCE of the Prim solar cell being as high as 21.33%.

image file: c8ee02575d-f6.tif
Fig. 6 Photovoltaic characteristics of series Prim PVK solar cells. (a) JV characteristics. (b) EQE quantum efficiencies. (c) TD and total energy losses of Prim cell and single unit MAPbI3 Cell.
Table 1 Photovoltaic parameters of serial Prim PVK cells using MAPbBr3, MAPbIBr2, MAPbI2Br, MAPbI3 active layers
Device Active layer d (nm) V oc (V) J sc (mA cm−2) J EQE (mA cm−2) FF PCEb (%)
a The standard deviation of PVK film thickness is ±5 nm. b Average PCE in brackets represent the standard deviation of 15 devices.
Cell 1 MAPbBr3 93 1.54 ± 0.03 5.38 ± 0.05 5.36 ± 0.02 0.76 ± 0.01 6.30 ± 0.02
Cell 2 MAPbIBr2 122 1.45 ± 0.03 5.43 ± 0.05 5.40 ± 0.03 0.77 ± 0.01 6.07 ± 0.02
Cell 3 MAPbI2Br 160 1.25 ± 0.03 5.53 ± 0.05 5.46 ± 0.03 0.78 ± 0.01 5.43 ± 0.02
Cell 4 MAPbI3 340 1.15 ± 0.03 5.42 ± 0.05 5.43 ± 0.02 0.77 ± 0.01 4.81 ± 0.02
Prim cells Cell 1–Cell 4 5.29 ± 0.05 5.35 ± 0.05 5.33 ± 0.03 0.75 ± 0.01 21.33 ± 0.02

To investigate the influence of the light incident angle on the efficiency of Prim PVSCs, the angle-dependent efficiency was measured, as shown in Fig. S7 (ESI); the PCE decreased when θ deviated from the optimal angle of −15°. When θ was increased to zero, more photons were absorbed by Cell 1, leading to a lower photocurrent in Cell 4; thus, the overall current density decreased. When θ decreased to −40°, the multi-reflected light could not totally cover the four subcells, leading to decreased photocurrent and efficiency due to outgoing photons. From the perspective of geometrical optics, the results show that the specific device construction parameters (i.e., device area of the subcell, the distance between the subcell and the reflector) codetermine the optimal incident light angle.

From the EQE in Fig. 6b, Cell 1 absorbs part of the solar light in the range of 350 nm to 550 nm, which leaves about 35% for Cell 2 in the range of 450 nm to 550 nm. The EQE of Cell 2 has a maximum value of 70% at 550 nm. Also, the EQE overlap between Cell 2 and Cell 3 in the range of 500 nm to 600 nm was observed. The EQE overlap between Cell 2 and Cell 3 is in the range of 600 nm to 720 nm; this EQE overlap resulted in some thermodynamic losses. Notably, the absorption band-edge of MAPbBr3 and MAPbI3 layers are 550 nm and 800 nm. So, it is impossible to divide the photocurrent equally for four subcells without EQE overlapping. In order to have equal Jsc for the four subcells, the active layer in Cell 1 should not absorb all photons in its absorption range. If there is complete absorption, the Jsc of Cell 1 will reach as high as 9 mA cm−2, and the following cells cannot reach that value. Therefore, this device can be further improved by introducing larger bandgap PVK materials, i.e., MAPbBr3−xClx (compared to MAPbBr3), as the active layer in the top cells. Therefore, more space to improve Voc and PCE based on MAPbBr3−xClx and MAPbI3−xBrx PVK films could be expected. Fig. S11 (ESI) provides preliminary calculations using this Prim structure. Assuming the ideal case with unchanged FF and no spectrum overlap between subcells, the factor of efficiency improvement could be estimated by dividing the average Voc of the series device by that of the single unit device. This shows that the PCE can be potentially improved by 40% using 6 series subcells (e.g. by dividing the solar spectrum from 350 nm to 800 nm into 6 spans of about 60 nm each) as compared to the single junction PVK cells. With the practical consideration of losses in FF and Jsc values, the highest PCE can still reach near 30% improvement based on the state-of-the-art 21% PCE of single unit PVSCs.60

To evaluate the energy loss rate ηloss of PVSCs, we propose the formula, according to the Shockley–Queisser model:15,42

image file: c8ee02575d-t1.tif(1)
ηloss(λ) = ηTD(λ) + (1 − EQE)(2)
where the first term of ηTD(λ) is thermodynamic loss rate of converted photons, the second term 1 − EQE(λ) is the energy loss rate of unconverted photons. EQEi(λ) is the external quantum efficiency of Cell i (i = 1, 2, 3 or 4) for the photons with a wavelength of λ, Voc,i is the open circuit voltage of Cell i, q is electron charge, h is Planck's constant, c is the speed of light. The detailed formula derivation can be seen in the ESI. The thermodynamic loss rate ηTD(λ) and overall energy loss rate ηloss(λ) versus photon wavelength λ of MAPbI3 and Prim PVSCs are shown in Fig. 6c. The MAPbI3 single device has a similar TD loss of 0.24 and total loss of 0.52 as compared to that of the Prim cell at the wavelength of 710 nm, whereas the MAPbI3 cell exhibits a much higher TD and total loss for converting high-energy photons. This suggests that the Prim cell reduced the TD loss and total loss by 18%, leading to the PCE enhancement from 18.6% to 21.3%. This improvement is quite high as compared to the reported tandem PVSC based on the MAPbBr3 and MAPbI3 subcells.44 Notably, this Prim cell has quite a high fabrication yield and is quite convenient for construction via individually preparing state-of-the-art subcells without worrying about the complicated multilayer stacking of tandem structure.

In conclusion, new prismatic PVSCs with light trapping configurations were designed by connecting perovskite devices with different bandgaps in the same horizontal plane, instead of the vertical multilayer stacking of tandem structures. The Prim PVSC device could guide the flow of light inside the device, thus, the four subcells based on MAPbBr3, MAPbIBr2, MAPbI2Br and MAPbI3 active layers could subsequently harvest high-to-low energy photons, generating Voc of 1.54 V, 1.45 V, 1.25 V and 1.15 V, respectively. As a result, the series Prim device generated a record Voc of 5.3 V, which showed a remarkable mitigation in the thermodynamic losses of the device. By carefully designing the absorption of four active layers, the photocurrents of four subcells were well matched, eventually leading to the high PCE, beyond 21%, for the Prim PVSC. This is the first report of a new PVSC device featuring thermodynamic loss mitigation and light trapping to achieve the highest Voc and performance, which could help to establish new device concepts for breaking the performance bottleneck of PVSCs. In a practical sense, Prim PVSC devices could also access flexible modules for large area application.


Fabrication of PVSCs

Before the fabrication of perovskite solar cells, the substrates were rinsed by sonication in detergent and deionized water, acetone, and isopropyl alcohol in sequence. After drying in a N2 stream, the substrates were further cleaned by plasma treatment for 40 s. Then, the 5 wt% Cu-doped NiO hole-transporting layer (HTL) was formed by spin-coating onto the substrates.29 MAPbI3−xBrx PVK layers were then evaporated onto the HTL in the four-source vacuum evaporation system. The thermal evaporation of the four precursors PbBr2, MABr, PbI2 and MAI was carried out in a thermal evaporation system. Four precursors were loaded in separate crucible heaters and the sample substrates were fixed on a rotatable substrate holder toward the precursor sources, where the positions of PbBr2 and MABr evaporation sources were close, as were the PbI2 and MAI sources. After the pressure of the evaporator chamber was pumped down to 10−7 mbar, MAI, MABr, PbI2 and PbBr2 were then heated to the set temperatures of 80 ± 10 °C, 70 ± 10 °C and 280 ± 15 °C, 210 ± 30 °C, respectively. When fabricating MAPbI3 layers, the deposition rates of MAI and PbI2 were both set at 0.2 and 0.2 Å s−1, respectively, to achieve a molar ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]1 for the two precursors, while the baffles of the MABr and PbBr2 sources were shut. For depositing the MAPbBr3 layer, the baffles of MAI and PbI2 sources were shut. When depositing the MAPbBrI2 layer, the deposition rates of MAI, PbI2, MABr and PbBr2 sources were set at 0.1, 0.1, 0.2 and 0.2 Å s−1, respectively. Similarly, the deposition rates of four precursors were set at 0.2, 0.2, 0.1 and 0.1 Å s−1 when depositing the MAPbBr2I layer. The temperature of the substrate holder was kept at 50 °C during the deposition. After that, the substrates are placed in a dish to form PVK films with better crystallinity via 100 μL DMF[thin space (1/6-em)]:[thin space (1/6-em)]CB (1[thin space (1/6-em)]:[thin space (1/6-em)]20 v/v) co-solvent vapor treatment for 10 min.52 Then, the PVK films were thermally annealed on the hotplate to evaporate the solvent residues. After that, solvent drops of the electron transporting layer (ETL) of ICBA (15 mg mL−1 in chloroform) and PCBM (15 mg mL−1 in chloroform) were dropped onto the film surface of MAPbBr3 and MAPbI3, respectively, and the nearby drops gradually came together during the evaporation of solvent on the hotplate. Thus, ETL on the MAPbBr3 and MAPbI3 active films were the pure ICBA and PCBM layer, respectively. ETL on the MAPbIBr2 and MAPbI2Br films were a mixture of ICBA and PCBM. After dip-coating ETLs, solvent annealing of the active layer was carried out by covering the NiO/MAPbI3−xBrx/ETL film with a petri dish for 24 hours in a glove box before the film was sent into the vacuum evaporator for depositing an electrode. The patterned 120 nm Al electrodes were evaporated through the mask onto the ETLs to form a connected Prim device, which formed a single unit device and electronic connected Prim device.

Characterization and measurements

The current density versus voltage (JV) characteristics of PVSCs were measured under N2 conditions using a Keithley 2400 source meter. A 300 W xenon arc solar simulator with an AM 1.5 global filter operated at 100 mW cm−2 was used to simulate the AM 1.5G solar irradiation. The illumination intensity was corrected by using a silicon photodiode with a protective KG5 filter. Masks were attached to define the effective area of the subcell of 1.05 × 4 mm2 for accurate measurement. The active device area is defined as the effective projection area of incident light into the Prim cell, which is 1.05[thin space (1/6-em)]cos(θ) × 4 mm2, where θ is the angle between the incident light and Prim cell; the width of 1.05 mm is the sum of the width of the subcell (1 mm) and the distance between subcells (0.05 mm). The Prim PVSC was placed at −15° to a horizontal line when measuring the light illuminating characteristics. The active device area was 4 ± 0.01 mm2. The EQE measuring system used a lock-in amplifier to record the short-circuit current under chopped monochromatic light from 350 nm to 850 nm. Theoretical Jsc values were obtained by integrating the product of the EQE spectrum with the AM 1.5G solar spectrum, which agreed with the measured Jsc to within 5%. The IQE spectra were calculated by the division of the EQE by the light absorption rate of the active layer.

Materials characterization

The structural analysis of MAPbI3−xBrx PVK films was conducted by X-ray diffraction (XRD) using Cu Kα radiation (λ = 0.1542 nm). The absorption spectra of MAPbI3−xBrx PVK films were obtained using a UV-Vis spectrometer with an integrated sphere. The photoluminescence spectra of PVK films were obtained using a spectroscopy system with laser excitation at 405 nm. The simulations of light distribution, light absorption efficiency and thickness dependence of the photocurrent of the individual layer within the devices were obtained using the transfer matrix method (TMM). The optical properties of each layer were represented by the index of refraction (ñ = n + ik), which were measured using a variable angle spectroscopic ellipsometer (VASE). All the theoretical simulations of optical properties are based on the assumptions of planar interfaces and isotropy for all layers within the device.

Conflicts of interest

There are no conflicts to declare.


This research was funded by the Foundation for Innovation Research Groups of the National Natural Science Foundation of China (NSFC) (Grant No. 61421002), National key research and development program (Grant No. 2018YFB0407100-02), the NSFC (Grant No. 6117703261675041 and 21372168), the Project of Science and Technology of Sichuan Province (Grant No. 2016HH0027). Dr Jiang Huang also thanks for the financial support of the Fundamental Research Funds for the Central Universities (Grant No. ZYGX2010Z004). C.-Z. Li thanks the financial support from NSFC (Grant No. 21674093 and 21722404) and the National 1000 Young Talents Program hosted by China.


  1. S. D. Stranks, G. E. Eperon, G. Grancini, C. Menelaou, M. J. P. Alcocer, T. Leijtens, L. M. Herz, A. Petrozza and H. J. Snaith, Science, 2013, 342, 341–344 CrossRef CAS PubMed .
  2. J. Huang, Y. Yuan, Y. Shao and Y. Yan, Nat. Rev. Mater., 2017, 2, 17042 CrossRef CAS .
  3. A. Kojima, K. Teshima, Y. Shirai and T. Miyasaka, J. Am. Chem. Soc., 2009, 131, 6050–6051 CrossRef CAS PubMed .
  4. B. R. C. E. National Renewable Energy Laboratory, http://www.nrel.gov/ncpv/images/efficiency_chart.jpg, 2018.
  5. Y. Deng, X. Zheng, Y. Bai, Q. Wang, J. Zhao and J. Huang, Nat. Energy, 2018, 3, 560–566 CrossRef CAS .
  6. W. S. Yang, B. W. Park, E. H. Jung, N. J. Jeon, Y. C. Kim, D. U. Lee, S. S. Shin, J. Seo, E. K. Kim and J. H. Noh, Science, 2017, 356, 1376–1379 CrossRef CAS PubMed .
  7. C. Momblona, L. Gil-Escrig, E. Bandiello, E. M. Hutter, M. Sessolo, K. Lederer, J. Blochwitz-Nimoth and H. J. Bolink, Energy Environ. Sci., 2016, 9, 3456–3463 RSC .
  8. J. Ávila, C. Momblona, P. P. Boix, M. Sessolo and H. J. Bolink, Joule, 2017, 1, 431–442 CrossRef .
  9. G. Longo, C. Momblona, M.-G. La-Placa, L. Gil-Escrig, M. Sessolo and H. J. Bolink, ACS Energy Lett., 2018, 3, 214–219 CrossRef CAS .
  10. H. Zhou, Q. Chen, G. Li, S. Luo, T.-B. Song, H.-S. Duan, Z. Hong, J. You, Y. Liu and Y. Yang, Science, 2014, 345, 542–546 CrossRef CAS PubMed .
  11. H. Chen, F. Ye, W. Tang, J. He, M. Yin, Y. Wang, F. Xie, E. Bi, X. Yang, M. Grätzel and L. Han, Nature, 2017, 550, 92 CAS .
  12. Y. Wu, X. Yang, W. Chen, Y. Yue, M. Cai, F. Xie, E. Bi, A. Islam and L. Han, Nat. Energy, 2016, 1, 16148 CrossRef CAS .
  13. H. Dong, Z. Wu, J. Xi, X. Xu, L. Zuo, T. Lei, X. Zhao, L. Zhang, X. Hou and K. Y. Jen, Adv. Funct. Mater., 2018, 28, 1704836 CrossRef .
  14. B. Yang, O. Dyck, J. Poplawsky, J. Keum, A. Puretzky, S. Das, I. Ivanov, C. Rouleau, G. Duscher and D. Geohegan, J. Am. Chem. Soc., 2015, 137, 9210 CrossRef CAS PubMed .
  15. W. S. Yang, J. H. Noh, N. J. Jeon, Y. C. Kim, S. Ryu, J. Seo and S. I. Seok, Science, 2015, 348, 1234–1237 CrossRef CAS PubMed .
  16. M. Saliba, T. Matsui, J. Y. Seo, K. Domanski, J. P. Correabaena, M. K. Nazeeruddin, S. M. Zakeeruddin, W. Tress, A. Abate and A. Hagfeldt, Energy Environ. Sci., 2016, 9, 1989 RSC .
  17. T. M. Koh, K. Fu, Y. Fang, S. Chen, T. C. Sum, N. Mathews, S. G. Mhaisalkar, P. P. Boix and T. Baikie, J. Phys. Chem. C, 2014, 118, 16458–16462 CrossRef CAS .
  18. N. Pellet, P. Gao, G. Gregori, T.-Y. Yang, M. K. Nazeeruddin, J. Maier and M. Graetzel, Angew. Chem., Int. Ed., 2014, 53, 3151–3157 CrossRef CAS PubMed .
  19. M. Saliba, T. Matsui, K. Domanski, J.-Y. Seo, A. Ummadisingu, S. M. Zakeeruddin, J.-P. Correa-Baena, W. R. Tress, A. Abate, A. Hagfeldt and M. Gratzel, Science, 2016, 354, 206–209 CrossRef CAS .
  20. D.-Y. Son, J.-W. Lee, Y. J. Choi, I.-H. Jang, S. Lee, P. J. Yoo, H. Shin, N. Ahn, M. Choi, D. Kim and N.-G. Park, Nat. Energy, 2016, 1, 16081 CrossRef CAS .
  21. J. H. Noh, S. H. Im, J. H. Heo, T. N. Mandal and S. I. Seok, Nano Lett., 2013, 13, 1764–1769 CrossRef CAS PubMed .
  22. T. J. Jacobsson, J.-P. Correa-Baen, M. Pazoki, M. Saliba, K. Schenk, M. Gratzel and A. Hagfeldt, Energy Environ. Sci., 2016, 9, 1706–1724 RSC .
  23. P. Docampo, J. M. Ball, M. Darwich, G. E. Eperon and H. J. Snaith, Nat. Commun., 2013, 4, 2761 CrossRef PubMed .
  24. J. Chen, L. Zuo, Y. Zhang, X. Lian, W. Fu, J. Yan, J. Li, G. Wu, C.-Z. Li and H. Chen, Adv. Energy Mater., 2018, 8, 1800438 CrossRef .
  25. K. Yan, J. Chen, H. Ju, F. Ding, H. Chen and C.-Z. Li, J. Mater. Chem. A, 2018, 6, 15495–15503 RSC .
  26. D. Zhao, Z. Zhu, M. Y. Kuo, C. C. Chueh and A. K. Y. Jen, Angew. Chem., Int. Ed., 2016, 55, 8999–9003 CrossRef CAS PubMed .
  27. Y. Lin, L. Shen, J. Dai, Y. Deng, Y. Wu, Y. Bai, X. Zheng, J. Wang, Y. Fang, H. Wei, W. Ma, X. C. Zeng, X. Zhan and J. Huang, Adv. Mater., 2017, 29, 1604545 CrossRef PubMed .
  28. Q. Jiang, L. Zhang, H. Wang, X. Yang, J. Meng, H. Liu, Z. Yin, J. Wu, X. Zhang and J. You, Nat. Energy, 2016, 2, 16177 CrossRef .
  29. J. H. Kim, P. W. Liang, S. T. Williams, N. Cho, C. C. Chueh, M. S. Glaz, D. S. Ginger and A. K. Jen, Adv. Mater., 2015, 27, 695–701 CrossRef CAS PubMed .
  30. C. Huang, W. Fu, C. Z. Li, Z. Zhang, W. Qiu, M. Shi, P. Heremans, A. Y. Jen and H. Chen, J. Am. Chem. Soc., 2016, 138, 2528–2531 CrossRef CAS .
  31. H. Chen, W. Fu, C. Huang, Z. Zhang, S. Li, F. Ding, M. Shi, C.-Z. Li, A. K.-Y. Jen and H. Chen, Adv. Energy Mater., 2017, 7, 1700012 CrossRef .
  32. F. Ullah, H. Chen and C.-Z. Li, Chin. Chem. Lett., 2017, 28, 503–511 CrossRef CAS .
  33. N. J. Jeon, H. Na, E. H. Jung, T.-Y. Yang, Y. G. Lee, G. Kim, H.-W. Shin, S. Il Seok, J. Lee and J. Seo, Nat. Energy, 2018, 3, 682–689 CrossRef CAS .
  34. J. Burschka, N. Pellet, S.-J. Moon, R. Humphry-Baker, P. Gao, M. K. Nazeeruddin and M. Graetzel, Nature, 2013, 499, 316–319 CrossRef CAS .
  35. Z. Zhang, W. Fu, H. Ding, H.-X. Ju, K. Yan, X. Zhang, F. Ding, C.-Z. Li and H. Chen, Adv. Mater. Interfaces, 2018, 5, 1800090 CrossRef .
  36. X. Yin, P. Chen, M. Que, Y. Xing, W. Que, C. Niu and J. Shao, ACS Nano, 2016, 10, 3630–3636 CrossRef CAS PubMed .
  37. F. Fan, F. Thomas, J. Timo, A. Enrico, B. Benjamin, Y. Songhak, B. Stephan and A. N. Tiwari, Nat. Commun., 2015, 6, 8932 CrossRef PubMed .
  38. S. Yang, W. Fu, Z. Zhang, H. Chen and C.-Z. Li, J. Mater. Chem. A, 2017, 5, 11462–11482 RSC .
  39. T. Leijtens, K. A. Bush, R. Prasanna and M. D. McGehee, Nat. Energy, 2018, 3, 828–838 CrossRef CAS .
  40. K. Gao, Z. Zhu, B. Xu, S. B. Jo, Y. Kan, X. Peng and A. K.-Y. Jen, Adv. Mater., 2017, 29, 1703980 CrossRef PubMed .
  41. S. Dong, Y. Liu, Z. Hong, E. Yao, P. Sun, L. Meng, Y. Lin, J. Huang, G. Li and Y. Yang, Nano Lett., 2017, 17, 5140–5147 CrossRef CAS PubMed .
  42. W. Shockley and H. J. Queisser, J. Appl. Phys., 1961, 32, 510–519 CrossRef CAS .
  43. A. Polman and H. A. Atwater, Nat. Mater., 2012, 11, 174–177 CrossRef CAS .
  44. J. H. Heo and S. H. Im, Adv. Mater., 2015, 28, 5121–5125 CrossRef .
  45. A. G. Imenes and D. R. Mills, Sol. Energy Mater. Sol. Cells, 2004, 84, 19–69 CrossRef CAS .
  46. H. Uzu, M. Ichikawa, M. Hino, K. Nakano, T. Meguro, J. L. Hernández, H.-S. Kim, N.-G. Park and K. Yamamoto, Appl. Phys. Lett., 2015, 106, 013506 CrossRef .
  47. K. Tvingstedt, V. Andersson, F. Zhang and O. Inganäs, Appl. Phys. Lett., 2007, 91, 123514 CrossRef .
  48. D. Zhao, Y. Yu, C. Wang, W. Liao, N. Shrestha, C. R. Grice, A. J. Cimaroli, L. Guan, R. J. Ellingson, K. Zhu, X. Zhao, R.-G. Xiong and Y. Yan, Nat. Energy, 2017, 2, 17018 CrossRef CAS .
  49. K. A. Bush, A. F. Palmstrom, Z. J. Yu, M. Boccard, R. Cheacharoen, J. P. Mailoa, D. P. Mcmeekin, R. L. Z. Hoye, C. D. Bailie and T. Leijtens, Nat. Energy, 2017, 2, 17009 CrossRef CAS .
  50. F. Sahli, J. Werner, B. A. Kamino, M. Bräuninger, R. Monnard, B. Paviet-Salomon, L. Barraud, L. Ding, J. J. Diaz Leon, D. Sacchetto, G. Cattaneo, M. Despeisse, M. Boccard, S. Nicolay, Q. Jeangros, B. Niesen and C. Ballif, Nat. Mater., 2018, 17, 820–826 CrossRef CAS .
  51. P. Löper, S.-J. Moon, S. M. de Nicolas, B. Niesen, M. Ledinsky, S. Nicolay, J. Bailat, J.-H. Yum, S. De Wolf and C. Ballif, Phys. Chem. Chem. Phys., 2015, 17, 1619–1629 RSC .
  52. H. Yu, X. Liu, Y. Xia, Q. Dong, K. Zhang, Z. Wang, Y. Zhou, B. Song and Y. Li, J. Mater. Chem. A, 2015, 4, 321–326 RSC .
  53. E. T. Hoke, D. J. Slotcavage, E. R. Dohner, A. R. Bowring, H. I. Karunadasa and M. D. Mcgehee, Chem. Sci., 2015, 6, 613 RSC .
  54. T. Duong, Y. L. Wu, H. Shen, J. Peng, X. Fu, D. Jacobs, E. C. Wang, C. K. Teng, K. C. Fong and M. Stocks, Adv. Energy Mater., 2017, 7, 1700228 CrossRef .
  55. B. Suarez, V. Gonzalez-Pedro, T. S. Ripolles, R. S. Sanchez, L. Otero and I. Mora-Sero, J. Phys. Lett., 2014, 5, 1628–1635 CAS .
  56. S. A. Kulkarni, T. Baikie, P. P. Boix, N. Yantara, N. Mathews and S. Mhaisalkar, J. Mater. Chem. A, 2014, 2, 9221–9225 RSC .
  57. Z. Zhu, Y. Bai, T. Zhang, Z. Liu, X. Long, Z. Wei, Z. Wang, L. Zhang, J. Wang and F. Yan, Angew. Chem., Int. Ed., 2014, 53, 12571 CAS .
  58. X. Zhu, D. Yang, R. Yang, B. Yang, Z. Yang, X. Ren, J. Zhang, J. Niu, J. Feng and S. F. Liu, Nanoscale, 2017, 9, 12316 RSC .
  59. J. Y. Jeng, K. C. Chen, T. Y. Chiang, P. Y. Lin, T. D. Tsai, Y. C. Chang, T. F. Guo, P. Chen, T. C. Wen and Y. J. Hsu, Adv. Mater., 2014, 26, 4107–4113 CrossRef CAS PubMed .
  60. M. Saliba, S. Orlandi, T. Matsui, S. Aghazada, M. Cavazzini, J.-P. Correa-Baena, P. Gao, R. Scopelliti, E. Mosconi, K.-H. Dahmen, F. De Angelis, A. Abate, A. Hagfeldt, G. Pozzi, M. Graetzel and M. K. Nazeeruddin, Nat. Energy, 2016, 1, 15017 CrossRef CAS .


Electronic supplementary information (ESI) available: Materials preparation, device fabrication and characterization, optical simulation data. See DOI: 10.1039/c8ee02575d

This journal is © The Royal Society of Chemistry 2019